📝 ESSAY
Three gates, three prices. The first should be nearly free.
Two organisations can run the same demand process on the same platform and get opposite pipelines. One is flooded with ideas nobody will sponsor, the other hears almost nothing from the business at all. The difference is rarely the intake form or the scoring model, it is who pays, and for what.
📍 IN BRIEF
The funding model is the first gate of platform demand. It sets a price on asking, and that price filters which ideas ever reach governance. Hold one price for an idea's whole life and you either flood the pipeline or empty it. The fix is to price the journey in stages, cheap to explore, charged to the unit that benefits to scale, shared to run.
The demand process gets redesigned every couple of years, with better forms and sharper scoring. The funding model underneath it is usually inherited from whoever signed the original licence, and it is often a primary driver in how the real decisions get made.
Organisations keep tuning the visible gate, sharper criteria, tighter business cases, a better board pack, and the shape of the pipeline barely moves. Flooded intakes stay flooded, silent intakes stay silent. If the demand process were genuinely deciding, then tuning it would change the outcome, and it does not, because the deciding happens earlier, in what it costs someone to ask.
The price of asking is set before the board sits
Every funding model puts a price on proposing work, and that price filters the pipeline before any reviewer sees it.
Watch what happens at the two extremes of that price. When the centre pays for everything, asking is free. A business unit that wants a new capability risks nothing by requesting it, so it requests everything, and the intake fills with ideas that sound worthy and belong to no one. The stated benefits inflate, because the person claiming them will never be asked to fund them. The board stops governing and starts queue-managing, and the loudest units win by volume.
Reverse the polarity and the failure inverts with it. When every request is charged straight back to the unit that raised it, asking is expensive, so units only raise what they can already defend to their own finance partner. Anything exploratory, anything whose value is genuinely uncertain, gets taxed out of existence at the door. The intake goes quiet, the platform looks disciplined and mature, and it is quietly stagnating, because the ideas that would have compounded were priced out before anyone with a portfolio view ever saw them.
Neither team chose these outcomes, because in most organisations the funding model was never designed at all. It is the fossil of a procurement decision, the shape of whichever budget happened to absorb the platform in year one. The demand board judges what arrives. The funding model decides what arrives, and it has been deciding since before the board existed.
Central, devolved and hybrid are filters, not accounting
Each of the three funding models removes a different kind of idea from your pipeline, so choose by what you can afford to lose.
Call this the price of asking. A funding model is a tariff on demand, and every tariff has a blind spot. Central funding, one pool governed from the middle, buys you consistency, a single roadmap and clean compliance, which is why it suits large estates in regulated industries. Its blind spot is ownership, because when nobody outside the centre pays, nobody outside the centre truly owns an outcome, and demand arrives as wishes rather than commitments.
Devolved funding, where each unit pays its own way, buys the opposite. Ownership is real, local priorities get met, and autonomous units move at their own pace. Its blind spot is the shared play, because the most valuable platform work, the common data foundation, the workflow that spans four departments, benefits everyone and is bought by no one. Under a devolved tariff those ideas do not lose the argument at the board, they are never proposed.
The hybrid model, a centrally funded core with unit-funded additions, is the standard recommendation for good reason, and it is where most mature platform organisations land. But naming the model is not the decision, the decision is where the line sits and who can move it. A hybrid whose boundary is renegotiated deal by deal is just a central budget with extra arguments. The line between what the centre funds and what units fund has to be a published rule, because the moment it becomes a negotiation, the price of asking becomes unpredictable, and an unpredictable price suppresses demand almost as effectively as a high one.
✅ PRINCIPLE
Pick the funding model by its blind spot, not its benefits. Central funding filters out ownership. Devolved funding filters out the shared play. A hybrid only works when the line between them is a published rule, not a negotiation.
Price the journey, not the idea
Make exploring cheap, make scaling pay, make running shared.
The deeper mistake sits under all three models, and it is holding one price for every stage of an idea's life. An idea that costs the same to explore as to scale will mostly not be explored. Venture investors solved this problem decades ago, and the answer was staging the money against the evidence. Seed money is small, quick and cheap precisely because the idea is unproven, and the serious pricing happens at the later rounds, against evidence the earlier round paid to create. Nobody prices the whole journey at the start, because at the start nobody knows what the journey is worth.
A platform tariff can work the same way. Exploration should be the cheapest thing in the system, a small, capped, centrally held pot that a team can draw on in days, with a hard ceiling per experiment so the downside is bounded. Scaling is where the price arrives, because once a pilot proves value, the unit that captures that value pays to scale it, and this is exactly where cost-recovery mechanics like usage-based charging or a flat platform levy belong, at the point where benefit is demonstrated, never at the door where it is still a guess. Running is shared, allocated by consumption, because a live capability serves everyone who uses it and its cost should follow that use.
Then the tariff itself needs a rhythm, because a price set once and left alone drifts out of contact with the portfolio it is supposed to steer. There is a well-established view, McKinsey have argued it for years, that budgets locked in annually and never reallocated against evolving outcomes work directly against iterative delivery. The remedy is not a bigger annual ask, it is a shorter re-pricing cycle. A cross-functional funding group, the same body that owns your investment portfolio, reviews the split each quarter, moves money from the stages that are starving to the stages that are queuing, and retires allocations that have stopped earning their place. Governance in this design does not approve individual ideas, it owns the tariff, and it prices the next quarter of asking.
Each funding model filters a different idea out of your pipeline. The staged hybrid moves the filter from purse to evidence.
Where this doesn't apply
A young platform with a single sponsor does not need a tariff. When one executive funds everything and the pipeline is eight ideas long, staging the money adds ceremony without adding signal, and the honest structure is one purse, well spent. Equally, some regulated environments mandate central financial control, and the funding model is not yours to choose. Even there the staging logic survives in miniature, because you can hold explore, scale and run as separate lines inside the central budget, and re-price them on the same rhythm, without moving a pound outside the centre.
The bottom line
Do not start by redesigning the intake process again. Start by finding out what an idea costs to propose in your organisation today, at each stage of its life, and who pays it. Then set three deliberate prices, exploring cheap and capped, scaling paid by the unit that benefits, running shared by consumption. Publish the line between central and unit funding as a rule, and re-price the whole tariff quarterly like the operating decision it is. Your demand board can only govern the ideas your funding model lets through, so govern the funding model first.
Show me who pays for your platform, and I will tell you what your innovation pipeline looks like.
P.S. The next time your intake volume moves sharply, ask what changed in the money before you ask what changed in appetite. A pipeline that floods or empties is usually responding to a price, not a mood.